Paul GrahamVerified account

A thing I think is underappreciated: prices measure changes which have often already happened in the world. The change in price is not itself usually the event itself. (Speaking of prices for assets here more than for products.)

For example, during the financial crisis, the value destruction wasn’t house prices falling, mortgages going under water, derivatives blowing up, or the bailout. These were all epiphenomena. The value destruction was society overallocating productive capacity to housing.

(Certain cities in the US are currently engaged in a contest to see who can destroy the most value by illegalizing deploying productive capacity for housing, an event which also impacts prices, in the other direction.)

Away from housing, one sees this in tech as well. Companies have to “grow into” valuations but the valuations themselves largely structurally lag metrics in the business which suggest that the firm is on a trajectory to possibly plausibly grow into those valuations.

Pricing attempting to simultaneously predict the future while being set in reference to the past is one of the things which causes so much confusion in the tech industry about it, because many folks’ intuitions about how companies should be valued are informed by regular life.

Regular life, including many great businesses, doesn’t have virtually any curves which look like 10% month over month growth sustained per year. So people largely are calibrated to assume that they mostly don’t happen. (They do happen.)

People just don't get exponential growth in general. I don't think it occurred much during human evolution. Sophisticated people know not to bet against 1.1^n, but I've never met anyone who could estimate it intuitively for arbitrary n.